Rec.709 ODT looks incorrect on consumer HDTVs about aces-dev HOT 5 CLOSED

I guess it depends on how good you think a "late-model consumer HDTV" is at correctly displaying images, or more specifically if they even attempt to model the conditions required for that specific CTL file.

That CTL is for a Rec 709/1886 display like a correctly calibrated 'HD Video' monitor (or my HDTV at home in 1886 mode in the appropriate viewing conditions).

Under the best interpretation of their actions the manufacturers are setting up their TVs in a way to best sell them under the typical environments they think they will be used, which is typically nothing like that of a typical mastering display, nor is it the exactly same as an sRGB monitor is specified to be under.

I think your requirement for matching the appearance of the two displays under side by side conditions is not something those CTFs are designed for, thus the images don't match

Kevin

from aces-dev.

Hi Kevin,

Thanks for the explanation. Maybe I misunderstood the comment in the ODT that the CTL could be used for HDTV mastering. I took this to mean mastering content for consumers, who typically own "consumer HDTVs", and didn't figure that it was for a calibration profile that would look significantly different from what most consumer's TVs (which are usually not professionally calibrated) would produce.

My goal in opening this issue was as much to gain understanding as reporting a perceived "bug". For context, the ACES RRT and Rec709 ODT has begun to see adoption in my industry (video game development) as a "filmic tonemapping" curve (see for example https://knarkowicz.wordpress.com/2016/01/06/aces-filmic-tone-mapping-curve/ ). At my company, we are in the process of integrating it into our engine (via a curve fit similar to the one in the article), and I would like to do it "correctly", or at least take a principled approach to any changes we make. It worries me that there seems to be such a noticeable discrepancy between the 709/1886 standards (which the ACES 709 ODT seems to follow, as best as I can tell) and what ordinary uncalibrated TVs actually produce.

From my point of view, it seems I have two options:

1. Use the ODT as-is, and accept that our game will look incorrect on most consumer's TVs, but will be correct on the minority of "correctly-calibrated" TVs
2. Modify the ODT to produce output that looks more correct on most consumer's TVs, but will look incorrect on calibrated TVs

Neither of these options are very appealing, but I think that from a pragmatic point of view, option 2 would win out, as uneasy as that makes me. I'm hoping that I've missed something, though!

Thanks,
Jasmin

from aces-dev.

HD Video should be viewed and quality assessed on a calibrated reference monitor under the correct viewing conditions.

The variability of TVs and viewing environments is too big to be able to cover.

from aces-dev.

from aces-dev.

Thanks for the feedback. After some investigation, I believe I can now explain what I was seeing.

• The PC monitors I've tested all exhibit non-standard gamma response, even when in "sRGB" mode. Specifically, the darker end of the curve has a gamma of about 2.6-3.0 depending on the monitor, while the lighter end of the curve was about 2.3-24.
• The TV I was using had a gamma curve closer to 2.2.
• The OOTF resulting from the Rec.709 OETF followed by the TV's 2.2 gamma curve is a very close match for the OOTF resulting from the sRGB OETF followed by the PC monitors' gamma curve, as shows by the plot below (sRGB axes, see here for an interactive version).

Thanks,
Jasmin

from aces-dev.